The proposed research seeks to add to our knowledge about the relationship between widespread social stress and the prevalence of mental illness. A major criticism of previous work in this area is that it has failed to sufficiently account for the impact of mental health service providers on this process. This work proposes to build upon previous work done by the Principal Investigator in this area with support from NIMH's Center for the Study of Metropolitan Problems. We have identified four variable domains which we argue affect observed mental illness in a significant manner. These are the level of stress in the catchment area, population, availability of services and the range of services offered. Using both single equation and two-stage least square regression models, we are proposing to model variables from these four domains as independent variables in an attempt to isolate the impact of the various domains upon variations in observed increases. The data base for this effort will build upon a 26-year series (1950-75) of mental health service utilization episodes for the Fitchburg Leomister, Massachusetts catchment area. The data set includes episodes of both inpatient and outpatient service utilization. As part of this research effort, we plan to expand the data base to include not just episodes of utilization but also measures of service availability. Additionally, we propose to extend the data series through 1979. The models developed will explore relationships between utilization, service availability, stress and referral, and disposition networks. To measure stress as a factor in utilization, we will use measures of employment and unemployment for the catchment area, refining the methodology originally developed by Brenner (1973) for studies of this type.